Biochemistry 1993, 32, 7861-7865
7861
Interaction and Influence of Phenylalanine- 198 and Threonine- 199 on Catalysis by Human Carbonic Anhydrase IIIt Xian Chen,t Chingkuang Tu,* Philip V. LoGrasso,*.s Philip J. Laipis,ll and David N. Silverman'.* Department of Pharmacology and Therapeutics and Department of Biochemistry and Molecular Biology, University of Florida, Gainesville, Florida 32610 Received March 19, 1993; Revised Manuscript Received May 4, 1993
Site-directed mutants of human carbonic anhydrase I11 were used to examine the role of Thr199 and its interaction with Phe-198 in the catalyzed hydration of COz. Threonine-I99 is a hydrogen bond acceptor for the zinc-bound water, and Phe-198 forms part of the hydrophobic side of the active-site cavity of carbonic anhydrase 111. Catalytic activity for a total of five single and double mutants at residues 198 and 199 was determined by stopped-flow spectrophotometry and l 8 0 exchange between C 0 2 and water Ala resulted in a 4-fold decrease in the measured by mass spectrometry. The replacement Thr-199 k,,/K, for hydration of C 0 2 . We tested the hypothesis that the 25-fold increase in the kat/& for hydration of C 0 2 accompanying the replacement Phe-198 Leu in isozyme I11 is caused by changes in the interaction of Thr-199 with the zinc-bound water or the transition state for catalysis. Comparison of hydration of C02 by the single and double mutants of isozyme 111 containing the replacements Thr-199 Ala and Phe-198 Leu was consistent with an interaction between these two sites. ABSTRACT:
-
-
Carbonic anhydrase I11 is found predominantly in skeletal muscle (Gros & Dodgson, 1988) and is the least catalytically efficient of the seven known isozymes of mammalian carbonic anhydrase (Tashian, 1989). Human carbonic anhydrase I11 (HCA 111)' has a maximal turnover which is 300-fold smaller than that for HCA I1 at physiological pH (Jewell et al., 1991; Engberg et al., 1985; Tu et al., 1983). Catalysis by isozyme I11 occurs in two separate stages and shares many common features with catalysis by isozyme I1 (Silverman & Lindskog, 1988). The first stage is the conversion of C02 into HCOs(eq 1) and is believed to occur by the direct nucleophilic attack of the zinc-bound hydroxide on COz (Silverman & Lindskog, 1988). The second stage is the proton-transfer steps which CO,
ki
kzIHz01
$
$
+ EZnOH- k-i EZnHC0,-
k-z
EZnH20
+ HC0,-
-
-
011.1 204
(1)
regenerate the zinc hydroxide form of the active site (eq 2). k3
EZnH,O
+ B z EZnOH- + BH'
(2)
k-3
Here B is the proton acceptor and can be a small buffer molecule (Tu et al., 1990; Paranawithana et al., 1990), water in the active-site cavity (Kararli & Silverman, 1985), or a side chain of the enzyme such as His-64 in the mutant K64H2 ?This work was supported by Grant GM25154 from the National Institutes of Health. * To whom correspondence should be addressed at Box 5-267, Health Center, University of Florida, Gainesville,FL 326 10-0267. Telephone: 904-392-3556. FAX: 904-392-9696. t Department of Pharmacology and Therapeutics. I Present address: Sandoz Research Institute, 59 Route 10, East Hanover, NJ 07936. 11 Department of Biochemistry and Molecular Biology. * Abbreviations: HCA 111, humancarbonic anhydrase111;Ches, 2-(Ncyc1ohexylamino)ethanesulfonic acid; Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; Mes, 2-(N-morpholino)ethanesulfonic acid; Mops, 3-(N-morpholino)propanesulfonic acid; Taps, 3- [ [tris(hydroxymethyl)methyl]amino] propanesulfonic acid.
0006-2960/93/0432-7861$04.00/0
FIGURE1: Residues near the zinc in bovine carbonic anhydrase I11 according to the crystal structure of Eriksson and Liljas (1 993). HCA I11 which then transfers the proton to buffer in solution (Jewell et al., 1991). The crystal structure of bovine CA I11 shows a backbone conformation very similar to that of HCA 11, with a root mean squared difference in backbone atoms of 0.92 A (Eriksson & Liljas, 1993; Eriksson et al., 1988; Eriksson, 1988), and the amino acid identity between these two enzymes is 56%. In both of these structures, the zincbound water is a hydrogen bond donor to the O y l of the side chain hydroxyl group of Thr-199 (Figure 1). This residue is invariant in all forms of carbonic anhydrase sequenced to date (Tashian, 1989),except for the plant carbonicanhydrases which have an entirely different primary structure [see, for example, Fawcett et al. 1990)]. In turn, the hydroxyl of Thr199 is a hydrogen bond donor to the carboxylate of Glu-106
* The single-letter amino acid abbreviations are used in which K64H HCA 111 refers to the mutant of human carbonic anhydrase 111 in which Lys-64 has been replaced with His. 0 1993 American Chemical Society
7862 Biochemistry, Vol. 32, No. 31, 1993 (Eriksson & Liljas, 1993). Merz (1990) has discussed the importance of the interaction between the hydrogen bond acceptor, the hydroxyl side chain of Thr-199, and the donor, the zinc-bound water or hydroxidein isozyme 11;its elimination through mutagenesis is expected to decrease catalysis. Phenylalanine-198 is unique to isozyme I11 and appears in all forms sequenced to date [human, bovine, equine, mouse, rat; see Eriksson (1988) and Eriksson & Liljas (1993) for references]. Phenylalanine- 198 is not a buried side chain but forms part of the hydrophobic side of the active-site cavity, with the CS of Phe-198 8.1 A from the zinc (Figure 1) and its phenyl ring hydrogen-bonded to water in the active site (Eriksson, 1988; Eriksson & Liljas, 1993). This residue is not believed to play a direct role in the catalytic pathway but does influence catalysis (LoGrasso et al., 1991; Silverman & Lindskog, 1988). The replacement of Phe-198 with Leu in HCA I11 enhanced the k,t/Km for hydration by 25-fold and increased the PKa of zinc-bound water by at least 1 PKa unit compared with wild-type (LoGrasso et al., 1991). The corresponding residue at position 198 in HCA I1 is Leu; thorough studies of the catalytic and structural consequences of amino acid substitution at this position have been reported (Krebs et al., 1993; Nair & Christianson, 1993). We have examined the function of Thr-199 in catalysis by HCA I11 using site-directed mutagenesis. The mutant T199A for the hydration of C 0 2 HCA I11 had kat and k,,/K, decreased 4-fold compared with wild-type HCA 111; the apparent PKa for catalysis increased from near 5 to 6.3. The mutant T199S had no change in kcat and apparent pKa with a 3-fold increase in kcat/&. For isozyme 111, this identifies the role of the hydroxyl group at position 199 as stabilizing the zinc-boundhydroxideand the transition state for catalysis. Our results also support the hypothesis that the increase in reactivity and pKa of zinc-bound water caused by the Leu in HCA I11 is related to the replacement Phe-198 interaction of Phe-198 with Thr-199.
-
METHODS Mutagenesis and Expression. Mutants of human carbonic anhydrase I11 were prepared using bacterial expression vectors optimized for efficient site-directed mutagenesis and protein synthesis as described by Tanhauser et al. (1992). Thevectors were derived from the T7 expression vectors of Studier et al. (1990) and contained a bacteriophage f l origin of replication for production of single-stranded DNA. Both single-site and cassette mutants were prepared with thesevectors. Expression depended on the mutant and ranged from 0.1 to 20 mg/L on the basis of kinetic measurements. All mutations were confirmed by DNA sequencing of the expression vector used to produce the mutant protein. Enzyme Purification. Variants of carbonic anhydrase I11 were purified by gel filtration (Ultrogel AcA 44, LKB) followed by anion-exchange chromatography (DEAE-Sephacel, Pharmacia) with minor modifications of the procedure of Tu et al. (1986). The resulting mutant enzymes were estimated to be at least 96% pure, determined by 0.1% SDS-12% polyacrylamidegel electrophoresiswith staining by Coomassie Brilliant Blue. The concentrations of HCA I11 and these mutants were determined from the molar absorptivity of 6.2 X lo4 M-l cm-l at 280 nm determined for bovine CA I11 (Engberg et al., 1985). The mutants of HCA I11 with the replacement Phe-198 Leu showed potent inhibition with ethoxzolamide (Ki = 2 X M). In this situation, we were able to confirm the concentration of enzyme to within 10% of that determined from the absorptivity by titration with
-
Chen et al. ethoxzolamide using a Henderson plot (Segel, 1975). Each of these variants was stable for many days when stored at 4 OC in solutions of 50 mM Tris, pH 8.4. Steady-StateMeasurements. The rate of hydration of C02 was determined by stopped-flowspectrophotometry (Applied PhotophysicsModel SF.l7MV), measuring the rateof change of absorbanceof a pH indicator (Khalifah, 1971). The bufferindicator pairs (with the wavelengths observed) were Mesl (PKa 6.1) and chlorophenolred (PKa 6.3,574 nm), Mops (PKa 7.2) and p-nitrophenol (PKa 7.1, 400 nm), Hepes (PKa 7.5) and phenol red (pKa7.5,557 nm), Taps (pKa8.4) and m-cresol purple (PKa 8.3, 578 nm), and Ches (PKa 9.3) and thymol blue (pKa 8.9,590 nm). Experiments were carried out at 25 OC with 50 mM buffer (unless otherwise noted), and the total ionic strength of the solution was maintained at a minimum of 0.1 M using Na2S04. Initial velocities of the hydrolysis of 4-nitrophenyl acetate were measured (Beckman DU 650 or DU 7 spectrophotometer) by the method of Verpoorte et al. (1967) in which the increase in absorbance was followed at 348 nm, the isosbestic point of nitrophenol and the conjugate nitrophenolate ion. Measurements were made at 25 OC, and the ionic strength was maintained at a minimum of 0.2 M with Na2S04. Solutions contained a 50 mM aliquot of one of the buffers used in the measurements of C02 hydration. For both hydration of C02 and hydrolysis of 4-nitrophenylacetate, kinetic constants were estimated from initial velocities using a least-squares method (Enzfitter, Elsevier-BIOSOFT). Oxygen-18 Exchange. The rates of exchange of l*O between C02 and water and of l 8 0 between 12C- and 13Ccontaining species of C02 were measured using an Extrel EXM-200 mass spectrometer or a Dycor M-100 gas analyzer with a membrane inlet probe (Silverman, 1982). Solutions were at 25 OC and contained 25 pM EDTA to complex potentially inhibitory metal ions. No buffers were used. The total ionic strength of the solution was maintained at 0.2 M with Na2S04. The l 8 0 method is useful because it measures the rate of interconversion of COz and HCO3- at chemical equilibrium, R1, as shown in eq 3. The substrate dependence of R1 is given HCOO"0-
+ EZnH20
+
EZn180H- + C 0 2 H,O
(3)
by Rt/[E] = kcateX[S]/(Ke# + [SI) in which kateXis a rate constant for maximal interconversion of COz and HCO3-, K e d is an apparent substrate binding constant, and [SIis the concentration of all species of C02 (Simonsson et al., 1979). Values of kateX/Kefffor the enzymes were determined by nonlinear least-squares fit of the above expression for R1 to the data for varying substrate concentration as in Figure 3 or by measurement of R1 at values of [SI much smaller than Kef?, In theory and in practice, k,tex/Kef? is equal to k,t/Km obtained by steady-state methods (Simonsson et al., 1979; Silverman, 1982). RESULTS Because of the low activity and lower expression levels for the two mutants containing Ala-199, a more complete set of data for kat/Km was obtained by l 8 0 exchange (Figure 2) which requires about 10-fold less enzyme than stopped-flow; these values were in agreement within 25% with values of kMt/Kmobtained by stopped-flow. The remaining values in Figure 2 were determined by stopped-flowas were the values for the mutants containing Ser-199. Kinetic Constants. Maximal values of kat/&, for the hydration of C02 catalyzed by HCA I11 and mutants are
Catalysis by Carbonic Anhydrase I11
Biochemistry, Vol. 32, No. 31, 1993 7863
f l . 7 1.
d
I
............ ....
'051
PH FIGUW2: Dependence on pH of kat/Kmfor the hydration of COz catalyzed by (A)HCA 111, (H) T199A HCA 111, ( 0 )F198L HCA 111, and (0)F198L-Tl99AHCA 111. Measurements were made by stopped-flow spectrophotometry (HCA I11 and F198L HCA 111) and 1 8 0 exchange (T199A and F198L-Tl99A HCA 111) at 25 OC using NazS04 to maintain the total ionic strength of the solution (see text). The stopped-flowexperimentsused 50 mM buffer as described in the text; the 180-exchange experiments used no buffers. a
o.9
Phe 198 ' Ale 199 7x
I
-0.1
io4
1
-
b
o.8
Phe 198 ' Ala 199 5 x 102
I
Leu 198 ' Thr 199 7 x 106
(wild t w ) Phe 198 ' Thr 199 2 x 10' 0'' -1.4
1
-
3 x 101 C
Phe 198 ' Ala 199 6.3
I
1
Leu 198 ' Ala 199 6.7
I
1
2.4
Leu l g 8 ' Ala 199
I
w's.'
1
2.7
8 x 104
0.5
3 x 10' -1.9
Leu 198 ' Ala 199
0.3
(wild type) Pho 198 ' Thr 199
Leu 198 ' Thr 199 2 x 104
-
W i d type) Phe 198 ' Thr 189 5.0
.o.6
Pho 198 ' Ser 199 8 x 105
I
-1.9
-
1
-0.6
Leu 198 ' Ser 199 2 x 107
Phe 198 ' Ser 199 2 x 105 -1.4
0
Leu
I
1
198 ' Ser 199 2 x 104
Phe 198 ' Ser 199 5.0 1.2
03
Leu 198 ' Thr 199 6.9
-1.4
I
i
Leu 108 ' Ser 199 5.9
FIGURE 3: (a) Comparisons of the maximal values of kat/Km (M-l s-1) for the hydration of COz at 25 OC catalyzed by variants of HCA I11 obtained by site-directed mutagenesis at positions 198 and 199 in the active-site cavity. Values of kat/Km appear beneath each designated mutant. Thevaluesadjacent to the arrows are the changes in free energy barriers (kcal/mol) for the catalysis corresponding to the designated mutations. Free energy changes were determined using AAG = -RTln[(k~,/Km)m,u/(kat/Km)m.tl]. (b) Comparisons of kat (s-1, given beneath each designated mutant) for the hydration of COz catalyzed by variants of HCA I11 obtained by site-directed mutagenesis. Values of kat are those observed in a pH-independent region at pH
